llvm dev - May 2006 - [LLVMdev] Re: Release 1.6 LLVM-Cfrontend build error on cygwin

Dear llvmdev,
     I am new to LLVM , but have a task on writing a LLVM backend to generate
some architecture specific assembly file.
	 Here is my cygwin build setting: GCC 3.4.4 , BIN UTILITY 2.15 ,and all other
packages of the right version listed on the LLVM Getting Started doc.
    
	 I have successfully built Release 1.6 LLVM. All the tools has been installed
there /usr/local/bin .
	 But I have problem with building LLVM-Cfrontend that comes with this Release.
	 CFrontend build stopped after these lines
	
--------------------------------------------------------------------------
gcc   -O2 -DIN_GCC  -DDEFAULT_TARGET_MACHINE=\"i686-pc-cygwin\"  -W
-Wall -Wwrit
e-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -fno
-common   -DHAVE_CONFIG_H -DGENERATOR_FILE  -o genattrtab.exe \
 genattrtab.o genautomata.o \
 rtl.o read-rtl.o bitmap.o ggc-none.o gensupport.o insn-conditions.o print-rtl1.
o errors.o \
 varray.o ../libiberty/libiberty.a -lm
/usr/lib/gcc/i686-pc-cygwin/3.4.4/../../../libm.a: could not read symbols: Archi
ve has no index; run ranlib to add one
collect2: ld returned 1 exit status
make[1]: *** [genattrtab.exe] Error 1
make[1]: Leaving directory `/cygdrive/d/Project/LLVM/cfrontend/build/gcc'
make: *** [all-gcc] Error 2
------------------------------------------------------------------------------

	Any advice on that? I have done quite a lot of online search about LLVM build
info  for several days, but no progress till now.   Blocked some what:-(

	Then I came up with another idea, can I work with the local built LLVM
infrastructure and a prebuilt cfronte
nd?

	So I downloaded an experimental llvm-cfrontend cygwin binary （3.4-LLVM
20030924）as well as a prebuilt llvm binary.

	But I cannot get a simple hello.c work with the combination of that cfrontend
binary and my local built llvm infrastructure. I used "llvm-gcc -S hello.c;
llvm-as hello.s; lli hello.s.bc;"
, but failed with    
	"ERROR: Program used external function 'printf' which could not be
resolved!   Aborted"

	But isn't any call to standard library  resolved in runtime by the local
VM,lli?
	

	Strange thing is when I try that with the combination of the  prebuilt
experimental cfrontend  and the prebuilt experimental llvm which was found in
the same place as the cfrontend binary, the hello.c can run this time although
with the warning info

"WARNING: Program called __main but was not linked to libcrtend.a.
This probably won't hurt anything unless the program is written in C++.
hello world"

	Deeply confused now. Anyone coming for help?
	


Best Regards,

 				

　　　　　　　　Chuck
　　　　　　　　chucking at sjtu.org
　　　　　　　　　　2006-05-14

On Sun, 14 May 2006, JinTao wrote:
> 	Deeply confused now. Anyone coming for help?
I don't really know anything about the cygwin port, but can you try LLVM 
1.7?  1.6 is quite old by now...

-Chris

-- 
http://nondot.org/sabre/
http://llvm.org/

Chris Lattner <sabre <at> nondot.org> writes:
> 
> On Sun, 14 May 2006, JinTao wrote:
> > 	Deeply confused now. Anyone coming for help?
> 
> I don't really know anything about the cygwin port, but can you try
LLVM
> 1.7?  1.6 is quite old by now...
> 
> -Chris
> 
I tried 1.7 version llvm, but the result is even worse, I cannot build llvm 
now.

-----------------------------------------------------------
make[2]: Entering directory
`/cygdrive/d/Project/LLVM/llvm-1.7/build/tools/llc'
llvm[2]: Compiling llc.cpp for Release build
llvm[2]: Linking Release executable llc (without symbols)
collect2: ld terminated with signal 11 [Segmentation fault], core dumped
/usr/lib/gcc/i686-pc-cygwin/3.4.4/../../../../i686-pc-cygwin/bin/ld: BFD 
2.16.91
 20050610 assertion fail /netrel/src/binutils-20050610-1/bfd/cofflink.c:1926
make[2]: *** [/cygdrive/d/Project/LLVM/llvm-1.7/build/Release/bin/llc.exe] 
Error
 1
make[2]: Leaving directory
`/cygdrive/d/Project/LLVM/llvm-1.7/build/tools/llc'
make[1]: *** [llc/.makeall] Error 2
make[1]: Leaving directory `/cygdrive/d/Project/LLVM/llvm-1.7/build/tools'
make: *** [all] Error 1
-----------------------------------------------
I shall skip llc and have another try.

Cygwin build is so slow on my machine.

Also, I tried building LLVM on redhat 9, but have problem even running 
configure. 

So I want to know your working setting there.

Hello, Chuck.

You wrote Tuesday, May 16, 2006, 6:03:14 PM:

C> /usr/lib/gcc/i686-pc-cygwin/3.4.4/../../../../i686-pc-cygwin/bin/ld: BFD
C> 2.16.91
C>  20050610 assertion fail
C> /netrel/src/binutils-20050610-1/bfd/cofflink.c:1926
C> make[2]: ***
This is a bug in binutils also seen in mingw32 build. :(

I've sent bug report to binutils bugzilla this weekend.
Hope, this will be fixed sometimes.

As for me, CVS checkout of 8 of May was compiled without
any problems (latest CVS snapshot triggers an assertion).

-- 
With best regards,
 Anton                          mailto:asl at math.spbu.ru

Tuesday, May 16, 2006 9:37:34 PM

Faculty of Mathematics & Mechanics, Saint-Petersburg State University

Hello, Chuck.

You wrote Tuesday, May 16, 2006, 6:03:14 PM:

C> /netrel/src/binutils-20050610-1/bfd/cofflink.c:1926
C> make[2]: ***
I've tracked this assertion some more deep. The problem is in
LLVMSelectionDAG.o file.

If I run the next lines, I've got and assertion:
ld -r -o LLVMSelectionDAG.o ScheduleDAGList.o ScheduleDAGRRList.o
ld -o llc LLVMSelectionDAG.o

If I remove one of the SheduleDAG* files from the full list of objects,
LLVMSelectionDAG.o is built from, the assertion is not longer
triggered.

Any ideas, why? I think, there is duplicate symbols somewhere, but I'm
not sure.

-- 
With best regards,
 Anton                          mailto:asl at math.spbu.ru

Tuesday, May 16, 2006 10:43:52 PM

Faculty of Mathematics & Mechanics, Saint-Petersburg State University

What does the assertion say?
On Tue, 2006-05-16 at 22:47 +0400, Anton Korobeynikov
wrote:
> Hello, Chuck.
> 
> You wrote Tuesday, May 16, 2006, 6:03:14 PM:
> 
> C> /netrel/src/binutils-20050610-1/bfd/cofflink.c:1926
> C> make[2]: ***
> I've tracked this assertion some more deep. The problem is in
> LLVMSelectionDAG.o file.
> 
> If I run the next lines, I've got and assertion:
> ld -r -o LLVMSelectionDAG.o ScheduleDAGList.o ScheduleDAGRRList.o
> ld -o llc LLVMSelectionDAG.o
> 
> If I remove one of the SheduleDAG* files from the full list of objects,
> LLVMSelectionDAG.o is built from, the assertion is not longer
> triggered.
> 
> Any ideas, why? I think, there is duplicate symbols somewhere, but I'm
> not sure.
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20060516/90340531/attachment.sig>

Hello, Reid.

You wrote Tuesday, May 16, 2006, 11:13:19 PM:

RS> What does the assertion say?
The same text, as in the original e-mail:

<=cut=>

/usr/lib/gcc/i686-pc-cygwin/3.4.4/../../../../i686-pc-cygwin/bin/ld:
BFD 2.16.91 20050610 assertion fail
/netrel/src/binutils-20050610-1/bfd/cofflink.c:1926

<=cut=>

This assertions is in big function named:

bfd_boolean
_bfd_coff_link_input_bfd (struct coff_final_link_info *finfo, bfd *input_bfd)

The text around assertion:

          h = NULL;
          if (*indexp < 0)
            {
              h = *sym_hash;

              /* The m68k-motorola-sysv assembler will sometimes
                 generate two symbols with the same name, but only one
                 will have aux entries.  */
              BFD_ASSERT (isymp->n_numaux == 0
                          || h->numaux == isymp->n_numaux);
            }

-- 
With best regards,
 Anton                          mailto:asl at math.spbu.ru

Tuesday, May 16, 2006 11:23:43 PM

Faculty of Mathematics & Mechanics, Saint-Petersburg State University

Hello, Chuck.

You wrote Tuesday, May 16, 2006, 6:03:14 PM:

C> I shall skip llc and have another try.
The cheap workaround is to use .a files instead of relinked .o. If you
want, I can send patch.

-- 
With best regards,
 Anton                          mailto:asl at math.spbu.ru

Thursday, May 18, 2006 7:22:54 PM

Faculty of Mathematics & Mechanics, Saint-Petersburg State University